Jump to content


 


Register a free account to unlock additional features at BleepingComputer.com
Welcome to BleepingComputer, a free community where people like yourself come together to discuss and learn how to use their computers. Using the site is easy and fun. As a guest, you can browse and view the various discussions in the forums, but can not create a new topic or reply to an existing one unless you are logged in. Other benefits of registering an account are subscribing to topics and forums, creating a blog, and having no ads shown anywhere on the site.


Click here to Register a free account now! or read our Welcome Guide to learn how to use this site.

Photo

How to compress MP3 voice recording files?


  • Please log in to reply
12 replies to this topic

#1 saluqi

saluqi

  • Members
  • 620 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:southern San Joaquin Valley, Calfornia
  • Local time:03:53 AM

Posted 14 January 2018 - 08:21 PM

Long years ago I used a program called "switch" from NCH Software to shrink voice recording files drastically without significant loss of clarity.  My recollection is the original MP3 recordings (from a Sony pocket voice recorder) ran about 10 MB per minute, and the converted files about 10 minutes per MB, in other words a 100-fold reduction in file size without noticeable loss of quality (and yes, I know you couldn't do this with nature or music recordings, but the human voice in conversation has limited bandwidth).

 

That was a good many years ago and now, needing to do the same thing again, I have forgotten how I did it.  Does anyone have words of wisdom here?  I suppose it's a matter of playing with the bit rate?  and/or converting to some other format such as WMA?



BC AdBot (Login to Remove)

 


#2 Platypus

Platypus

  • Moderator
  • 14,972 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:08:53 PM

Posted 15 January 2018 - 05:23 AM

Maybe I'm missing something, but if Switch can do what you need, why not use it?

Incidentally, around 10MB per minute is CD standard, not MP3. Even so, 100-fold reduction in file size without noticeable loss of quality would be quite an achievement.

Edited by Platypus, 15 January 2018 - 06:37 AM.

Top 5 things that never get done:

1.


#3 Rocky Bennett

Rocky Bennett

  • Members
  • 2,755 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:New Mexico, USA
  • Local time:04:53 AM

Posted 16 January 2018 - 05:20 PM

Let me see, 28,000 bits per second divided by 100 = 280 bits per second. So you are saying that you can create an audio file with just 280 bits per second and still maintain audio quality. I don't believe it.


594965_zpsp5exvyzm.png


#4 Chris Cosgrove

Chris Cosgrove

  • Moderator
  • 6,904 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Scotland
  • Local time:11:53 AM

Posted 17 January 2018 - 07:10 PM

Any time I want to produce MP3s I use LameXP as a converter since, except for a memory stick for use in my wife's car, I keep my music as WAV files. LameXP gives a range of compression ratios from, I think, 320kBs down to about 40. Certainly in a relatively noisy environment like a car 320kBs is quite acceptable quality. The bottom end is a bit more like the PA system in your local railway station - on a bad day !

 

Chris Cosgrove



#5 Rocky Bennett

Rocky Bennett

  • Members
  • 2,755 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:New Mexico, USA
  • Local time:04:53 AM

Posted 17 January 2018 - 09:45 PM

Any time I want to produce MP3s I use LameXP as a converter since, except for a memory stick for use in my wife's car, I keep my music as WAV files. LameXP gives a range of compression ratios from, I think, 320kBs down to about 40. Certainly in a relatively noisy environment like a car 320kBs is quite acceptable quality. The bottom end is a bit more like the PA system in your local railway station - on a bad day !

 

Chris Cosgrove

 

 

I agree with everything you said.


594965_zpsp5exvyzm.png


#6 mightywiz

mightywiz

  • Members
  • 810 posts
  • OFFLINE
  •  
  • Gender:Male
  • Local time:03:53 AM

Posted 19 January 2018 - 06:24 PM

 

Any time I want to produce MP3s I use LameXP as a converter since, except for a memory stick for use in my wife's car, I keep my music as WAV files. LameXP gives a range of compression ratios from, I think, 320kBs down to about 40. Certainly in a relatively noisy environment like a car 320kBs is quite acceptable quality. The bottom end is a bit more like the PA system in your local railway station - on a bad day !

 

Chris Cosgrove

 

 

I agree with everything you said.

 

the average person cannot tell the difference between 320k and 128K, when you go less then 128K then the audio distortions get really noticable.   192K is CD quality and unless your a sound engineer or in the music industry you won't even be able to tell the difference from 192K compared to 320K.  I do audio editing and when I do key changes in the music if my audio is less then 320K I start getting noticeable sound distortions because of the audio compression and loosing to much data.



#7 Rocky Bennett

Rocky Bennett

  • Members
  • 2,755 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:New Mexico, USA
  • Local time:04:53 AM

Posted 19 January 2018 - 07:37 PM

 

 

Any time I want to produce MP3s I use LameXP as a converter since, except for a memory stick for use in my wife's car, I keep my music as WAV files. LameXP gives a range of compression ratios from, I think, 320kBs down to about 40. Certainly in a relatively noisy environment like a car 320kBs is quite acceptable quality. The bottom end is a bit more like the PA system in your local railway station - on a bad day !

 

Chris Cosgrove

 

 

I agree with everything you said.

 

the average person cannot tell the difference between 320k and 128K, when you go less then 128K then the audio distortions get really noticable.   192K is CD quality and unless your a sound engineer or in the music industry you won't even be able to tell the difference from 192K compared to 320K.  I do audio editing and when I do key changes in the music if my audio is less then 320K I start getting noticeable sound distortions because of the audio compression and loosing to much data.

 

 

 

This is funny because I have over 6,000 CDs and they are all 1,411 kbps, not 192 kbps.


594965_zpsp5exvyzm.png


#8 Chris Cosgrove

Chris Cosgrove

  • Moderator
  • 6,904 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Scotland
  • Local time:11:53 AM

Posted 19 January 2018 - 07:46 PM

C'mon guys - let's cut down on the quoting !

 

I know, I know, it's my pet hobby horse - but it is almost never necessary to quote a complete post, especially when it is the one immediately above the quote you are making. If you don't cut back I will go into my posting record and start quoting all the times i have brought this up - and that would make a lengthy post !

 

Chris Cosgrove



#9 nickos

nickos

  • Members
  • 207 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:cyprus
  • Local time:01:53 PM

Posted 20 January 2018 - 03:48 AM

good morning to all,

 

you can lower the overal bot rate to something lower than 64 kbps for speech only also if you encode to mono you an save some more speace

 

what about the quality ?

 

if your source is quite good and if the files are ''speech'' only you can encode with good results even with as low as 32 kbps

 

i don't know what quality offer all those audio converter (software) but if you encode your files under a pro sound editor like wavelab, adobe audition, sony sound forge (now magix) your sound quality will be good enough

 

you have to pay attention on the sound level (loud enough without clipping - a good level is -0.3 to -1.4 and if you think the sound level is still too low you can use a sound maximizer)

 

also, you have to eq your speech (cut any high frequencies above 4.500 khz and enhanced some mid frequencies between 1.200 to 2.85 with some bass cut as well)

 

with those settings your speech files even with a 32 kbps will be sounded quite quite while the mp3 size will remain low

 

good luck my friend



#10 mightywiz

mightywiz

  • Members
  • 810 posts
  • OFFLINE
  •  
  • Gender:Male
  • Local time:03:53 AM

Posted 23 January 2018 - 10:48 AM

and yes cd's are recorded @ 1,411kbps WAV format uncompressed.  WAV & MP3 are two different beasts.  1411kbs WAV recording are an uncompressed RAW format.



#11 saluqi

saluqi
  • Topic Starter

  • Members
  • 620 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:southern San Joaquin Valley, Calfornia
  • Local time:03:53 AM

Posted 03 February 2018 - 04:18 PM

Hm, perhaps I will just have to play around and see what I can get.  I am using a Sony ICD PX-333 pocket voice recorder - a far cry indeed from the Uher and Nagra tape recorders and Sennheiser mikes, parabolic reflectors etc. I used to lug around, decades ago, to make "nature" recordings in Africa etc.  Lots of water over the dam since then, and all I'm looking for in this particular situation is clear intelligibility of conversational speech in the compressed file.  Speaker to mike distance no more than 10 feet.  Speech volume as in addressing a not too large room with say a dozen people.

 

I've had Switch and WavePad on my system for years, but just recently updated them (well, new installs of the most recent versions) and now Secunia is telling me that the LAME encoder has to be updated manually.  I'd never seen that mentioned (and certainly not by Secunia PSI) until recently, and searching among my files does not find anything called LAME, so do I suppose that this is built into one or both of the NCH programs mentioned?

 

Edit - these MP3 files were recorded at a sampling rate of 44.1 kHz and a bit rate of 192 kb/sec.


Edited by saluqi, 03 February 2018 - 05:56 PM.


#12 saluqi

saluqi
  • Topic Starter

  • Members
  • 620 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:southern San Joaquin Valley, Calfornia
  • Local time:03:53 AM

Posted 03 February 2018 - 08:34 PM

A few experiments today:

 

The original MP3 file, bit rate 192 kb/sec: duration 43 minutes 19 seconds, file size 59.4 MB so MP3 compression to roughly 15% of original file size

 

Converted at bit rate 24, quite intelligible, a bit "boomy", file size 7.43 MB

Converted at bit rate 16, still intelligible with a bit of effort, file size 4.95 MB

Converted at bit rate 8, still mostly intelligible but quality noticeably worse, file size 2.47 MB

 

there's quite a bit of background noise (ventilation equipment etc.) in the original recording, and it gets worse at the lower bit rates.  I haven't yet applied any filtering, etc., though I'm sure it would help.

 

I notice the sound quality of the reduced files is considerably affected by what software I use to play them back.

I converted them to MP3 format, might some other format be preferable?



#13 nickos

nickos

  • Members
  • 207 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:cyprus
  • Local time:01:53 PM

Posted 04 February 2018 - 04:32 AM

good morning to all,

 

if you want to encode to mp3 speech with small file size and relative good sound quality just follow my advice in my earliest post

 

but now i read that you have recorded natural effects and speeches as well so i would suggested the follow

 

if you want to keep those sound effects just encode your mp3 at the highest possible bit rate (320 kbps cbr)

 

natural sound effects have a very broad frequency response and dynamic range better to encode to wave 44.1 khz @ 24 bits !

 

with mp3 you gonna loose all those precious harmonics that the mp3 standards remove 

 

if you want to keep your speeches but without those effects (background noises etc) then you have to go for the sony's spectralayers software, the only one program for the moment that can remove effectively unwanted background noise and other elements from a recording

 

you cannot remove effectively by filtering ''only'' any background and unwanted noises, you can only reduce it's sound level within the recording and furthermore by filtering your recording will be dull, certainly you can add equalization but again the results will not be fully satisfactory (for my ears)

 


Edited by nickos, 04 February 2018 - 04:37 AM.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users